Christel de Bruijn, PhD @christeldebruijn.bsky.soc statsepi Precisely. It's literally one population at *baseline*. The only difference is the coin flip. Therefore any p<0.05 on baseline characteristics is by definition a Type I error since it's one population.

validated doesn't mean valid validated doesn't mean good validated doesn't mean useful validated doesn't mean approved validated doesn't mean sufficient validated doesn't mean accurate validated doesn't mean best available validated means we did a test

Yes, gut bacteria do produce some vitamins and other essential components, but the amounts are really small. Don't rely on your gut bacteria (only 200g, remember) to supply you with nutrients! Hannelore Daniel

Important take-home messages from Hannelore Daniel. To move microbiome research forward, and avoid unfounded hype, there are many issues that must be considered.

While genomics explain ~3% of the variation in BMI, the metabolomic component explains ~50%! -Jerzy Adamski on the potential of metabolomics i health research.

Should we all adopt the DASH diet? Check out our editorial in #EJPC, commenting on a recent paper by Daniel Ibsen and colleagues. academic.oup.com/eurjpc/advance…

Twitter friends! Help support our new study. Feeding the future study (FEED) is looking for UK residents to complete a one-off 20 minute online survey about their eating habits, dietary motivations and lifestyle. Please share widely🙏. SHIFT WHO/Europe Oxford Population Health (OxPop) ceu.ox.ac.uk/research/feedi…

When does non-differential #misclassification *NOT* cause bias towards the null, and why should we avoid over-reliance on #heuristics? This paper was super fun to write up with Matthew Fox @AKWesselink and Tim Lash! Now out in Am J Epidemiology #epitwitter academic.oup.com/aje/advance-ar…

Recent reviewer nonsense: "Add p-values, as this is for a clinical audience" "Non-linearity is confusing; re-do analyses at clinically meaningful cut-points" "Something's wrong if complex stats methods are needed" "Reduce methods section - it's distracting for clinical readers"

Our response to the recent modeling study estimating the effect of sustained dietary changes on life expectancy. @DierkesJutta @tho_ols osf.io/5wdfp/

Yes, please do this!

Medical Research Bingo

Vitamin D is a case study in the challenges of nutrition assessment & of applying the principles of evidence-based medicine to nutrition, & how little citing a role of a nutrient in the body informs upon how much to ingest. It's definitely a lucrative biohacker brand builder tho.

During my PhD, I got alot of unhelpful advice Now I'm a PhD supervisor, I know 💯 what I needed to hear 4 things every PhD student should know 🧵 #phdchat thoughtsofaphd PhD_Genie The PhD Story Happy PhD Supervisor PhDiva Refocused

August reminder for meta-analysis folk: I-squared estimates between-study heterogeneity❌ I-squared is a test for between-study heterogeneity❌ I-squared > 50% means large heterogeneity❌ I-squared measures proportion of total variability due to between-study heterogeneity✅

New paper out on dairy products and cardiovascular disease outcomes, with (among others) @AntheaVanparys @DierkesJutta @tho_ols Kathrine Vinknes Åslaug Matre Anfinsen Kirsten Holven academic.oup.com/eurjpc/advance…

More than 25 million people use Google Docs. But most of us still don't know its best features. Here are 11 Google Docs tips that'll save you 100s of hours of your life:

If you're at the point where you're spreading fear about folks eating fruits in smoothie form, just stop. The hypothetical micro-optimization focus of so much health information focuses on a possibly existing tree while missing the forest.

This is the key challenge for personalised medicine wannabes. How do you distinguish between A) This works for X% all the time B) This works for everyone X% of the time C) Some mixture of A & B? (Hint: parallel group trial + responder” analysis doesn’t work.)